Correlation Analysis of Sleep Study Variables in Obese v. Non-obese Military Personnel Diagnosed with Obstructive Sleep Apnea

Abstract

Background: Several epidemiologic studies have found an association between obesity and the risk of Obstructive Sleep Apnea (OSA). However, the exact mechanisms involved in OSA among obese individuals and the interactions between obesity and OSA are unknown. One way to identify possible mechanisms of OSA specific to obesity is to explore differences in the pathophysiology of OSA between obese and non-obese patients. The military presents a unique opportunity to study non-obese patients because of the high number of non-obese soldiers that undergo sleep studies. Therefore, military sleep study records offer a prime opportunity to differentiate how OSA behaves in patients with and without obesity. Methods: Our cross-sectional study compared associations between diagnostic markers of OSA and selected physiologic sleep disturbances among obese (body mass index (BMI) ≥ 30 kg/ m2) and non-obese (BMI < 30 kg/ m2) patients with OSA. We reviewed and analyzed a database of OSA cases (N = 342) comprised of soldiers diagnosed with OSA who underwent polysomnography (PSG) at a major military medical center in 2010. The cases were divided into obese (n = 176) and non-obese (n = 166). Pearson correlations (r) were calculated for Apnea Hypopnea Index (AHI) among patients with mild OSA (5 ≤ AHI < 15), moderate/severe OSA (AHI ≥ 15), and both groups combined. PSG variables (arousal index (AI), minimum oxygen saturation (O2 Sat)) were compared between the obese and non-obese groups. Results: A statistically significant correlation between AHI and AI was only seen in moderate-severe OSA for obese patients (r = 0.55, p < 0.01) while significant correlations were seen in both mild (r = 0.20, p = 0.02) and moderate-severe OSA (r = 0.45, p < 0.01) for non-obese patients. On the other hand, statistically significant correlations between AHI and min O2 were found in both mild (r = -0.37, p < 0.01) and moderate-severe (r = -0.38, p < 0.01) OSA for obese patients while the correlation was only found in mild (r = -0.26, p = 0.01) OSA for non-obese patients. Furthermore, the correlation between AHI (≥ 5, without disease severity stratification) and min O2 sat was significantly (more than twofold) stronger for obese (r = -0.56, p < .01) than for non-obese patients (r = -0.27, p < 0.01) and the difference was also statistically significant (p < 0.01). Conclusions: Results suggest that obstructive events during sleep in non-obese and obese patients with OSA might involve different cascades of pathophysiologic events. For obese, the correlation between frequency of arousal (AI) and number of obstructive events (AHI) increases as the disease progresses. And the depth of hypoxia was consistently associated with the number of obstructive events throughout the disease severity. However, for non-obese, the correlations between frequency of arousal and number of obstructive events also increases as the disease gets more severe, but, the non-obese lost the association between the number of obstructive events and the depth of hypoxia as disease gets worse. Given that physiologic disturbances from the primary pathologic event of OSA may differ as a function of obesity, future study should focus on clarifying the different clinical manifestations of OSA in obese and non-obese patients and may further consider comparing the efficacy of treatment(s) for OSA between obese and non-obese patient groups.